GREGORY: FISH SKULLS 127 



forward displacement of the quadrate-articular joint would also give increased importance 

 to the lower part of the preopercular in its function of stiffening the posterior side of the 

 quadrate. The same movement would give space for the small postorbitals of a form like 

 Oxygnathus to grow downward behind the first circumorbital row. This postorbital series 

 is of course not to be confused with the true postorbital of the tetrapods, which is equivalent 

 merely with a single one of the inner or circumorbital series of fishes. 



Many of the Semionotidae tended to parallel the Sparidae, or breams, among teleosts 

 in the development of pebble-like teeth, presumably adapted to crushing mollusc shells. 

 Some of the more normal members of this series gave rise to the very long-lived and success- 

 ful genus Lepidotus (Fig. 22B). Others, becoming still more specialized, gave rise to the 

 peculiar Pyncodontids, with rows of pebble-like teeth clustered on a median cylinder in 

 the upper jaw and on the inner sides of the lower jaw. The skull pattern of these forms 

 became aberrantly modified, chiefly through consolidation and strengthening of its parts, 

 but as the' family has nothing to do with the line of ascent to later families, it may here 

 be passed by. 



Lepidosteids. — The garpike (Fig. 21 A, 24) of the existing ganoids preserves the basic 

 heritage of the semionotid skull, including the large circumorbital plates and the forwardly 

 inclined suspensorium, but it has become specialized in the great elongation of the snout, 

 in the fragmentation of the maxilla into numerous plates, and in many other features. 

 Goodrich (1909, pp. 342-344) has shown that in many features the garpikes agree with 

 Lepidotus, and suggests that very possibly Lepidosteus is merely a specialized late remnant 

 of the family Semionotidae. 



Recent authorities differ as to the identification of the bone beneath the opercular. 

 Tate Regan (1923a) identifies it as an interopercular, partly on account of its relations 

 with both the subopercular and the angular of the mandible. Mayhew (1924), on the other 

 hand, identifies this bone as the preopercular because it carries part of the operculo- 

 mandibular sensory canal. He also applies the term interopercular to the small bone that 

 lies behind and beneath the quadrate. This is also the identification of these elements 

 adopted by Goodrich (1909, p. 342). 



A dried skeleton of Lepidosteus (Figs. 23, 24) with all these elements in their nearly 

 natural positions shows without doubt that the large curved bone called by Regan the 

 interopercular has all the proper connections for the preopercular. Thus it articulates 

 with the lateral surface of the hyomandibular, running to the top of that element and 

 receiving the preopercular branch of the lateral line system. Then it runs downward 

 and forward, expanding greatly on its posterior and inferior borders so as to afford the main 

 outer brace for the suspensorium. This part of the bone evidently overlapped and then 

 thrust itself between the subopercular and the interopercular. The dorso-medial surface 

 of the ascending branch of the bone evidently aflfords a secure origin for the lateral parts 

 of the adductor mandibulae muscle. The inner border of this area has a contact with the 

 head of the interhyal, as has the preopercular of Salmo. Continuing forward, the bone 

 in question forms also the lateral brace for the lower end of the symplectic and for the 

 palatal extension of the metapterygoid. At its antero-inferior end it overlaps the true 

 interopercular, which as usual is fastened by ligament to the derm-angular. One reason 

 why the bone in question does not overlap the lateral posterior border of the quadrate in 

 the normal manner of a preopercular is that the latter is too small and has been crowded 



